Every year, millions of acres of forest and rangeland are burned in prescribed burns as well as wildfires. The costs associated with wildfires may be some of the largest we face as a society both in material goods and in life. The importance of managing fire fuels has increased with the development of the wildland-urban interface. With this increased emphasis has come the development of tools to assess, map, and simulate fuel maps at a landscape level. These fuel maps are then input into computer-aided wildfire simulation models that are used by land managers in the planning process. A current challenge for land managers is to find efficient ways to measure the amount and structure of fire fuels on a landscape level. Fuel models are one of the required inputs for software that mathematically computes wildfire rate of spread. Various methods have been used to develop fuel maps. It is the objective of this thesis to develop a method by which fuel models can be predicted and mapped on a landscape level through utilization of remotely sensed data. The proposed process for this method is: 1) develop landcover classification, 2) assess data analysis approaches for use in creation of predictive regression models, 3) correlation of data results to Natural Fuels Photo Series, and 4) translate Natural Fuels Photo Series classifications into fuel models described by Scott and Burgan.